Forum Discussion

Hanzo581's avatar
Hanzo581
Level 4
11 years ago

Finally need to fire up our Disaster Recovery Site

Since I have taken over Netbackup at my work we've never put any focus on our DR site and never even thought about really starting it up.

Well, a new command has taken us over and our new bosses want us to test our DR capabilities.  I've been looking through some documentation but I need real world advice.

Our DR environment is VASTLY different than our local production environment.

 

Here we have locally:

Master Server: Server 2008 R2

Media Servers: 7 (3 Unix, 3 Linux, 1 Windows)

Virtual Tape Library: Quantum DXi8500 (60 drives)

Physical Tape Library Quantum i2000 (12 drives)

 

Basically we write all to virtual and then use SLPs to make physical tapes.

 

What we have at DR:

Master Server : 2008 R2 (same NB version as prod)

Media Servers: 1 (Linux)

Virtual Tape Library : NONE

Physical Tape Library : Quantum i500 (1 Drive)

 

Basically, can you point me to the best possible documentation for anything pertaining to recovering netbackup at an offsite environment that is very different than the source.  Any and all info would be fantastic.

  • Hello,

    I would start with: A Guide to Site Disaster Recovery Options Doc (TECH66060)

    Also chapter 5 of the NetBackup Troubleshooting guide (DOC5180). 

    It sounds like your off to a great start on the DR site. The Master server is on the same version as the Production site and it has a tape library for a catalog recovery. 

    Is there any specific questions you have about it?

10 Replies

  • All day and no responses.   Guess no one wants to touch this one with a ten foot pole...

  • Hello,

    I would start with: A Guide to Site Disaster Recovery Options Doc (TECH66060)

    Also chapter 5 of the NetBackup Troubleshooting guide (DOC5180). 

    It sounds like your off to a great start on the DR site. The Master server is on the same version as the Production site and it has a tape library for a catalog recovery. 

    Is there any specific questions you have about it?

  • Not really specific questions, more of just a, "were starting pretty much from scratch" situation looking for some guidance.

     

    We already send tapes up to DR, we are going to start sending a catalog backup up with them...once we import the catalog, can we just simply restore whatever is on those DR tapes to any clients up there?  Do the media servers have to match or will any do when it comes to restoring?

  • You might not like this...but you need the powers that be to state the business requirements:rpo and rto - thats what you need to aim for. It might be as simple as "we have to DR everything", but probably not. We have hundreds of servers and apps but there are 3+ classes: warm DR (no netbackup at all), cold DR and no DR.

    And within warm and cold we have priorities ie what to do first. No DR is clearly no recovery whatsoever, typically dev environments.

    So then you know what servers to DR and what the rpo and rto are: rpo=where in the life of the server you have to recover to eg: 24hr (hence daily backup requirement), 7 day (hence weekly), 1 hr ( getting trickier...) and rto=time it takes to get services recovered to rpo: eg 2days. So everything you have to sign off on has metrics attached.This will determine how you back up and what resources you need at DR and will lead you to figure out the effort / resources to hit the rpo and rto. Then you go back to the business and tell em the news, good or otherwise.

    It doesnt matter if its radically different, what matters is it has to fit the bill, down to how each server is to be recovered. You've made no mention of servers at DR ie what the restore is to. Or the location, logistics, manpower, access, power , a/c, networking, ...

    I guarantee you'll find problems which will alter the way you do things currently.

    I am available for hire. Really.

    Jim

     

  • As it sits we really only backup production data, so everything we do has a DR copy made and sent to our DR location.  This is done weekly.  Right now I just set it up to where a catalog goes along with these weekly tapes.  I know we have a mirror of our environment (as far as destination client servers) go up there, so we would be restoring to them, though the names would be different.

    My God I wish the Government wasn't so cheap, would be great to actually high people to show us how to do this stuff instead of just pointing us to Google and saying have fun.

  • You are correct, the idea is to do a cataolg recovery at the DR site to get current backup records. From there you now have the image catalog and can do redirected restores to other servers and recover your data from any new tapes. 

    The DR media server could be different and you will want to set it up for restore failover. In the recovery guide it talks about going to NetBackup Management > Host Properties > Double-click on master server > Restore Failover and entering in the production media server and the failover media server used in your DR. (HOWTO33210)

  • So how you do it is pretty clear! Recover the catalog backup in DR, recover each of the servers.

    Write down all the servers and then how each will be recovered...from the ground up (BMR?) or via a restore and modify or maybe the servers are pre-built and you know what data to restor? If BMR then you'll need that at DR too. Sounds like you have your work cut out.

    Jim

     

  • Alrighty, so we have some test tapes and our catalog up there ready to import.  What I'd like to know, is if there is any chance of interference with our production environment here when we import the catalog up at our disaster recovery site?  We've never done this before so we're trying to sort out all the details.

    So basically is there any harm in our current environment being live when we fire up our DR netbackup environment and import the catalog?

     

    Thanks for any info.

  • You should be separte from your production so you would be fine. I assume your DR site is not connected to the production site?

  • As Terry says, thats a networking issue. Simple issue to rule out any chance of interference: hook up a console at the DR site master, login via console , physically remove any and all other network connections to that master server where you are doing the work, kick off the catalog recovery.

    Jim